A major cause of reduced efficiency in semiconductor devices such as rectifiers is inadequate cooling during normal operation. Two ways to alleviate this problem involve the use of a heat sink and a reduction in the wall thickness of the housing that encapsulates the semiconductor device.
FIGS. 1a and 1b are perspective and side views, respectively, of a Vishay® Semiconductor brand single phase inline bridge rectifier device 10 having four semiconductor die inside (not shown), manufactured by Vishay Intertechnology, Inc. Device 10 is through-hole mountable via leads 14, and includes an exterior epoxy housing 12 that protects the semiconductor dies—during operation of device 10, heat generated by the semiconductor dies is transferred through leads 14 and housing 12. FIG. 1B is a side view of device 10, illustrating the through-hole mounting thereof to a substrate 11 (a circuit board, for example), and further illustrating how a heat sink (such as a finned aluminum plate) 13 may be used to increase the thermal dissipation performance of housing 12—heat generated by dies within housing 12 is transferred to substrate 11 via leads 14, and through housing 12 to heat sink 13 and/or the ambient environment. Device 10, substrate 11, and heat sink 13 are cooled using a cooling technique such as natural or forced-air convection. The thermal conductivity of epoxy housing 12 is much less than that of heat sink 13, however, which often results in device 10 having poor thermal dissipation performance. Unfortunately, the housing 12 is necessary to protect the semiconductor device against moisture as well as assembly process and other environmental contaminants. The housing 12 is commonly formed by encapsulating the semiconductor device in a mold compound, such as a thermosetting plastic, applied by a transfer molding process.
In a typical transfer molding machine used in the semiconductors industry, a thin electronic workpiece mounted on a lead frame is clamped between two halves of a split mold. The mold defines a mold cavity around the device with sufficient clearance to allow mold compound to be injected and flow around the device to encapsulate it. During the molding process mold compound is injected into an inlet and air inside the mold escapes from a vent. A plunger drives the liquefied mold compound into the mold cavity. The mold compound is allowed cure and the mold is opened, releasing the encapsulated semiconductor device.
As previously mentioned, to increase heat dissipation device manufacturers would like to reduce the thickness of the encapsulating layer of the mold compound which encases each device. Thinner encapsulating layers also aid in improving device performance or reliability with regard, resistance to coating damage under thermal stress and other parameters. Another reason thinner encapsulating layers are desired is that in general smaller semiconductor devices are preferred over larger device. However, as the distance between the inner mold surfaces and the electronic workpiece is decreased, it becomes more difficult to obtain a high quality void-free encapsulant around the entire device.
To obtain a void-free seal, the liquefied mold compound must enter the mold inlet and entirely fill the space in the mold cavity before the mold compound flow front arrives at the mold vent. If the mold compound reaches the vent before the mold is completely filled, an air bubble is trapped in the mold, creating a void.
To completely fill the mold cavity, the mold compound must flow between the upper mold surface and the upper surface of the device, between the lower mold surface and the lower surface of the device, and into the space surrounding the outer perimeter of the device. However, as the distance between the upper and lower mold surfaces and the device is reduced, so as to make the encapsulating coating thinner, it becomes more difficult for the mold compound to penetrate these regions.
If this distance is reduced too far, the mold compound will flow around the outer perimeter of the device before the mold compound flow front has displaced the air in the space above and below the device. The result is a void in the encapsulation material as an air bubble is pinched off in the center of the device.
As a result, transfer molding of semiconductor devices with conventional equipment has required that the distance from the inner mold surfaces to the device be at least about 200-250 micrometers. This ensures that there will be laminar flow of the molding compound into the mold and around the device. The exact minimum distance limit is, of course, a function of the specific mold compound used, the fillers it contains and process parameters, such as temperature, but, in general, reducing the distance from the inner mold surfaces to the device to less than some minimum distance results in unacceptable manufacturing losses due to the formation of voids.
FIG. 2 is a cross-sectional view of a semiconductor device 10 such as depicted in FIGS. 1a and 1b. Similar to the device in FIG. 1, device 10 is through-hole mountable via leads 112, and includes an exterior epoxy housing 110 that protects the semiconductor dies 106—during operation of device 10, heat generated by the semiconductor dies 106 is transferred through leads 112 and housing 110. The thermal conductivity of epoxy housing 110 often results in device 10 having poor thermal dissipation performance. A major cause of reduced efficiency in semiconductor devices such as rectifiers is inadequate cooling during normal operation. Unfortunately, as discussed above, if the thickness of the housing is reduced to achieve better thermal conduction, molding failures such as IPE or voids 130 (see FIG. 2) tend to increase, leading to problems such as a failure of a high-potential (hipot) test or electric strength test, which is caused by a breakdown in insulation.
Accordingly, it would be desirable to accommodate both a heat sink and a thinner housing wall in order to adequately dissipate her in a semiconductor device such as a bridge rectifier.